Hawks, Doves, and Owls is the report of the “Avoiding Nuclear War” project carried out at the John F. Kennedy School of Government at Harvard. Some of its contributors have had direct experience of policy-making circles in Washington. What is more, its editors, Graham Allison, Albert Carnesale, and Joseph Nye, Jr., all of Harvard, have a clear message. The protection of US values and institutions is absolutely dependent on the avoidance of nuclear war. And since no one can know how such a war might be set off or, if started, controlled, that imperative need is also dependent on the avoidance of any war in which the Western Alliance and Warsaw Pact powers might become embroiled. What is more, the three editors recognize that “the nuclear threat creates a solidarity of interest between the two superpowers—against a total war in which they would be the greatest victims.”

It is against this background, and on the sensible assumption that “US-Soviet rivalry, superpower interest in Europe, and Third World instability” are unlikely to change in the near future, that the six scholars who contribute the specialized essays in this book consider how things could go wrong. The editors provide the introductory chapter, which sets out what they call the “agenda,” and they end the book with recommendations for action.

The first chapter, on accidental nuclear war, is by Paul Bracken of Yale. Whatever might have been the situation in years gone by, he argues, technological and organizational developments have made it all but inconceivable that a war could erupt today as a result of the sort of accident that is often discussed—for example because a nuclear weapon had been launched after an operator had incorrectly interpreted radar blips made by a flock of geese, or because of a failed computer chip, or even through the action of a “crazed military officer.” There are many who are not as optimistic, among them Daniel Ford, whose book I shall be discussing later.

Bracken is more worried about trouble arising from maritime accidents in which ships or submarines are sunk because of a hasty and incorrectly drawn conclusion that they were attacked as part of a larger plan. What is needed is a continuous channel of communication between the American and Russian authorities. For the rest, he wants the US to concentrate its “energy on preventing confrontations,” and the establishment of “design principles and rules of the road” to deal with periods of intense crisis. He is also worried because he has a hunch that Britain, France, and China “have given even less thought to these issues than the United States has.” If what he writes is an indication of how far the US has got in formulating “rules of the road” to deal with the danger of unintentional nuclear war, the three countries he mentions will not have far to go before they find themselves up against a brick wall.

In the essay that follows, Richard Betts, formerly a staff member of the Senate Select Committee on Intelligence, and now of the Brookings Institution, deals with surprise attack and preemption. From him we learn that the Russians keep their forces “on lower levels of day-to-day alert” than does the US. This is also stated in a later chapter by Stephen Meyer, who deals specifically with “Soviet Perspectives on the Paths to Nuclear War.” Betts rates very low the possibility that either side would embark upon a preventive war, since in a nuclear age such a step might well be suicidal. The only circumstances in which he can conceive of the US risking a preemptive attack would be when a state of tension had come about because political changes in the world order were moving decisively in the Russian favor, and when the US had a chance of limiting the scale of nuclear retaliation by means of a first strike at the Russian missile silos.

Here, however, is the problem. It is inconceivable that the US would launch a preemptive surprise attack unless it was certain that nuclear war was inevitable. But who would say that it was? “A cool and rational decision to this effect by a collective body is hard to imagine.” Both sides deploy a “triad” of nuclear forces, including missiles based on land, in the air, and under the sea. Since this in effect means that both possess an invulnerable second-strike force, the price of a preemptive strike would be the high risk of national suicide—regardless of who had attacked first. In order “to ensure consistency between political aims and military signals,” both sides, Betts argues, therefore clearly need to be “thoroughly educated in the dynamics of military alerts and crisis deployments.” And in this respect, he implies, there is greater coherence today on the Soviet than on the American side.


In effect this theme is taken up by Fen Hampson, who deals with the problem of maintaining peace in Western Europe. A possible “scenario” of how this could be broken is outlined by the editors in their introduction. An uprising might take place in East Germany which the authorities could not suppress; Russian troops would be called in to help; West Germans would infiltrate to help their brothers in the East; the Russians would block all the roads to West Berlin; a NATO supply airlift would be resisted by the Russians; and so on until there is full-scale land war with the risk of nuclear escalation.

Hampson considers other possibilities—for example, that there might be “a general loosening of the Western alliance that created opportunities for Soviet adventurism, or a deterioration in Eastern Europe that tempted the United States to reverse the Yalta accords, or, more likely, some combination of the two.” War could spread to Europe from the Middle East. There might also be changes in the balance of forces, particularly in the credibility of the US commitment to launch its strategic nuclear forces in the presumed defense of Europe.

When he considers the balance of forces, Hampson sensibly associates himself with those who, now that both sides have a superfluity of warheads, are not impressed by what he calls “bean counts” and what others refer to as “nuclear accountancy.” He is also concerned about the instability that would result from futile efforts to devise defenses against ballistic missiles. But the most important and immediate danger is that a major disruption of the political status quo could lead to hostilities between the NATO and Warsaw Pact powers, hostilities which, once started, could move on to nuclear disaster. If we are to continue to avoid war in Europe, there should never be any ambiguity about NATO’s intentions of a kind that could invite a preemptive attack by Warsaw Pact forces. For example, tactical nuclear weapons should not be deployed in a way that invites preemption or accidental launch. NATO’s conventional forces should always be adequate to deal with those they face. Equally important is a sensible political strategy to govern NATO in its dealings with the Warsaw Pact powers.

It is interesting that the Stockholm International Peace Research Institute has recently put out a book arguing along the same lines as does Fen Hampson.1 It is also worth noting that a recent study by Daniel Frei, carried out in cooperation with the United Nations Institute for Disarmament Research,2 agrees that the chances of an unauthorized nuclear war are “practically zero” but warns that one could start unintentionally, not only as escalation from a conventional war but because of a multitude of causes, all compounded in different degrees of “misjudgment, miscalculation and misunderstanding.”

Francis Fukuyama’s chapter, “Escalation in the Middle East and Persian Gulf,” parallels Hampson’s on Europe insofar as it provides a balanced account of the dangerous turns that events could take in the vast and troubled region stretching from Morocco in the west to Pakistan in the east. Ever since the end of World War II and the breakup of the old British and French colonial and imperial order, and ever since the foundation of the state of Israel, there has rarely, if ever, been a moment when two or more of the countries concerned were not at one another’s throats. The stranglehold on Western economies of the Middle East OPEC countries has also become a major disruptive issue, in which the interests of the US and the Soviet Union conflict. Both, however, have an obvious interest in avoiding hostilities, and Mr. Fukuyama rightly insists that the “single most important factor” influencing the outcome of past crises has been “political prudence on the part of national leaders”—hence the importance of cultivating “friendly talk” and better communications between the US and the USSR.

The likelihood of the superpowers being drawn into a catalytic nuclear war through their opposing interests in some dispute involving a third country, or because of the nefarious activities of some terrorist organization, is not rated highly by Henry Rowen in his contribution to Hawks, Doves, and Owls. Hetoo sees a clear community of interest between the two superpowers. But as Stephen Meyer, a consultant to the CIA and the Defense Department, remarks in another essay, it is not always easy to understand the official Soviet views about nuclear matters. As he puts it, we have no Russian memoirs on the subject, no transcripts of public governmental hearings, no talkative officials who have moved out of Moscow, as they do out of Washington, into industry or a university.

Meyer, a professional student of Soviet affairs, has little faith in publicly proclaimed Soviet military doctrine and policy, or in pronouncements clearly made for propaganda purposes. He does not mention it, but the way Khrushchev made it his business at the beginning of the Sixties to mislead the US about the missile strength of the USSR is a case in point. In order to discern what can be learned about “Soviet perspectives on the paths to nuclear war,” he prefers to extract what meaning he can from such published material as is available about Russian military plans, force structure, and deployment, and from an analysis of Soviet diplomatic behavior. He reminds his readers that “throughout the 1950s Soviet leaders watched as tens, then hundreds, of American ‘strategic’ delivery systems were deployed around the Soviet periphery,” with the not surprising consequence that the Russians are convinced that the US is committed to their destruction. Whichever way hostilities could begin, the Russians still fear a nuclear bolt from the blue.


Meyer also reminds us that, despite this fear, Soviet nuclear forces are always at a lower state of alert than are the American, and that Soviet nuclear weapons are not entrusted to the military as a whole, but are in the charge of an elite and specialist body which he believes has some connection with the KGB. Today only Russian ICBMs are on day-to-day alert and loaded with nuclear warheads, ready to retaliate against a US first strike. “No Soviet strategic bombers are on alert,” while “only a comparatively small fraction of Soviet SLBMs [submarine launched ballistic missiles] are available at any given time for second-strike retaliation under surprise attack conditions.” “Despite popular impressions,” Mr. Meyer continues—and one has to suppose that as a CIA and Defense Department consultant he should know—“most of the Soviet military machine is not geared up to go to war at a moment’s notice.” Hence Soviet military doctrine implies “other possible paths to nuclear war.”

His view is that the Russian leaders now believe that nuclear hostilities could not occur except as a result of some major crisis and confrontation or as an escalation of conventional wars. Any form of “NATO nuclear use” would “unleash a massive and devastating Soviet theater nuclear strike against all NATO military facilities.” This would escalate to global nuclear war only if the Soviet homeland were hit by American weapons, say Pershing IIs or cruise missiles based on the Western European mainland. Even then, the Politburo, not the military, would be the ones who would authorize a retaliatory strike. In a recent article in Foreign Affairs Albert Wohlstetter also argues forcefully that the Politburo keeps “the use of nuclear weapons under continuing control.”3 What matters therefore is the close political understanding that exists between the Politburo and the high military command. Be this as it may, Dr. Meyer warns that no Westerner could ever predict how the Soviets would react to a nuclear strike, from wherever it came. I suppose the Russians are equally in the dark about how the West would respond to an attack they might launch.

Although Hawks, Doves, and Owls is based on long and lengthy collaboration, the two essays in which the editors sum up its findings are a bit of a letdown. The preface tells us that the book is addressed first to “members of the policymaking community,” next to interested citizens, and third to “experts on nuclear weapons and international security.” The stylized presentation of a list of “dos and don’ts” about how to avoid nuclear war might well help the interested reader to organize his thoughts on the subject. But I am doubtful whether such a list, plus a few simple tables and diagrams, could ever be taken seriously either by political leaders or by their military advisers.

For example, do they need to be warned that it is wise to maintain a credible nuclear deterrent given that one is deployed by their potential enemy; or that cities cannot be defended against nuclear attack; or that conventional forces should be strengthened; or that the Soviet Union should not be provoked; or that the West, i.e., the US, should plan for ending a war if one begins; or that the US should “expect the unexpected”; or that nuclear weapons should not be treated as other weapons? These homilies are merely statements of the obvious, and they hardly do justice to the other essays in Hawks, Doves, and Owls. It would have been a different matter if the book had specifically explained when a nuclear force becomes “incredible,” or when “enough is enough,” or what degree of so-called modernization can be under-taken without destabilizing the present state of mutual deterrence. But such matters are left untouched.

Perhaps the project was launched with altogether too ambitious an aim. How could there be a textbook that described what should or should not be done, given this or that political crisis, as a physician might resort to a computerized data bank of the signs and symptoms of disease? It is inconceivable that any such book would have been consulted in the Cuban crisis of 1962, or would be were such a crisis to recur. The story of how President Kennedy and his close advisers conducted themselves during that period of confrontation is well known and was in fact the subject of a book by Graham Allison, one of the editors of Hawks, Doves, and Owls. I do not know what ordinary people in the United States believed was going to happen, but in Europe fear of a nuclear war was widespread. Moreover, no one knew what to do about it.

The British were kept closely informed about the progress of events, particularly as the UK ambassador to Washington, the late Lord Harlech (David Ormsby-Gore at the time), was a close friend of the President. On the assumption that the “nuclear balloon” was bound to go up when the President’s ultimatum expired, the top officials of the British defense establishment were summoned to a meeting at 9 AM on Sunday, October 28. Just as they sat down, the news came on the radio that Khrushchev had agreed to President Kennedy’s demands, including the removal of the missiles and the dismantling of the missile sites.

We looked at one another in relief. We had no contingency plans. Had hostilities broken out in the Caribbean, there was a possibility that the Russians would have immediately imposed a new blockade of Berlin, or taken some other action that would have brought war to Europe. But we had no plans. I remember Lord Mountbatten, chairman of the British Chiefs, then saying, “What would we have done if the Russians had not backed down? We must work this out.” There was no answer. I am sure that were a similar situation to arise tomorrow, it too would not be dealt with, whether in Washington, Moscow, or Brussels, according to a synthesis and analysis of past decisions and past judgments. Military crises do not spring full-blown onto the center of the stage in accordance with a given pattern. Suez, Cuba, Korea, Vietnam, the Falklands, Lebanon—each was different, each built up differently, and each could have evolved and ended differently from the way it did.

Even if one were to assume that guidelines and rules of conduct of the kind spelled out in the final chapter of Hawks, Doves, and Owls could have practical value, a necessary condition would be that Paul Bracken is right when he assures us that we need not worry about malfunctions of a mechanical kind in the nuclear system, or about some sergeant or major starting a nuclear war in a fit of madness. For if the system could, as it were, operate on its own—or fail to operate when it was commanded to do so—there would be an even more unpredictable set of circumstances to worry about in addition to the political and other possible upsets the book lays before us. And to Daniel Ford, the control of the nuclear system is nothing near as reliable as Bracken would have us believe.

A New Yorker journalist, Mr. Ford sets out his reasons for this view in The Button, which, is based on lengthy series of interviews, mostly with unnamed people who presumably were military personnel and should not have told him all they did. We learn that the Defense Department and the Joint Chiefs are largely dependent on commercial communications networks in maintaining their worldwide military command and control. Relay stations are in vulnerable places, and some critical parts of the early warning system against a missile attack have been “off the air” for some time because they are waiting for new computers or for thermionic valves to be replaced by modern solid-state devices.

A computer at NORAD (North American Aerospace Defense Command), the nerve center of the US defense system in Cheyenne Mountain near Colorado Springs, does not work. The one-star general who was showing Mr. Ford around explained what each of five telephones on the NORAD commander’s battle-station desk was for, but when the obliging officer picked up the telephone connecting directly with the White House, there was no one at the other end. Dr. John Steinbruner of the Brookings Institution, and once a member of the official world of Washington, told Mr. Ford that “given the extreme vulnerability of the entire [command and control] mechanism to the preemptive moves of the opponent,” the American strategic nuclear force was “primed to go—massively.”

This and other bits of information lead Mr. Ford to conclude that a command system which is “unready to deliver orders for retaliation but streamlined to convey commands for a first strike unmistakably suggests the basic intent of its designers.” So far as he is concerned, “the stated [US] national policy of retaliation” is merely “official rhetoric.” If the president had to decide without delay whether to launch or not to launch a nuclear strike against the USSR, he could not rely on getting together with his closest advisers, the secretaries of state and defense. Mr. Ford’s authority for telling us this is Henry Kissinger, whom he quotes as saying that the only way US strategy could be implemented would be by delegating authority to some field commander “who must be given discretion so that when he thinks a nuclear war has started, he can retaliate.”

If this is indeed the situation, if even half of what Daniel Ford tells us about the inadequacies of the US command and control system is true, the Avoiding Nuclear War project of the Kennedy School either overlooked some central questions or did not treat them in sufficient depth.

First, can a 100-percent trust ever be placed in any command and control system on which the fate of a nation could rest, and particularly on one which depends overwhelmingly upon electrical and electronic relays? It takes time—decades—to construct even the kind of early warning system against a missile attack that we believe we have now. During that time technology changes, so that elements which were introduced when the system was designed need to be updated and replaced as later components are added. Ford tells us the information from current US warning systems is transmitted to 1960s computers, which were built to analyze a strike of some fifty warheads. If they had to deal with a strike involving thousands of missiles or decoys, they would be overwhelmed with data and could assess neither the size of the attack nor the likely targets. Much the same could be said of our anti-aircraft defenses. Were that not the case, why should there now be mention, in discussions of President Reagan’s strategic defense initiative (SDI), of the need to update the anti-aircraft defense systems that started to be built in the Fifties?

Those who have the responsibility of translating into hardware the President’s concept of a defense in space against ballistic missiles know how formidable are the technological hurdles that need to be overcome. As the President would have learned from an official panel which was set up in 1983 under Mr. Fred Hoffman to consider the strategic implications of his proposed SDI, an imperfect defense, one which could only deal with a small but not a full-scale attack, would be a provocative step.4 Even if a defensive space astrodome were possible, it could not prevent nuclear weapons getting beneath it or past it. And we know that in the world of abstractions with which strategic analysts deal, the idea of controlling escalation, even if one assumed a perfect command and control system, is nonsense unless the two opposing sides were to cooperate closely. In any event, if an early warning system as well as all its related systems did work, and a nuclear launch was signaled, it would be too late. Nuclear war would have begun.

Because of his concern that an all-out nuclear exchange could result not only in mutual suicide, but also in the kind of global catastrophe now called nuclear winter, Wohlstetter, in the article he has just published in Foreign Affairs, sets out a scenario in which the two superpowers come to some kind of agreement restricting the use of nuclear weapons to military targets only, so as to avoid extensive civilian deaths and to keep the conflict limited. This would be fine—so far as it goes, and given that the actual military world resembled the one that armchair strategists invent so easily, and if there were mutual agreement, indeed international agreement, about what constituted military targets.

The reality is that no military leader, no one who has carried the responsibility of command in the European theater, can genuinely believe that he knows what would happen once one side had fired a nuclear weapon. The initiative would then pass immediately to the opposing side, regardless of what may be written in the tactical handbooks that assume the use of nuclear weapons (both sides have to have such handbooks to accord even abstract meaning or credibility to the deployment of such weapons). Moreover, the belief that nuclear weapons could be used in battle carries with it the implication that the authority vested in the supreme command for their use would have to be quickly delegated to army commanders, corps commanders, and even to divisional commanders. Such authority would have to be delegated right down to the level where the tide of battle had moved in such a way that, at least in theory, what happened next could be influenced by one or more nuclear shots delivered at targets at some distance from one’s own troops. Once this happened, there could be no central control. To the best of my knowledge, every nuclear war game that has been played out realistically on the basis of actual military dispositions in the European theater has shown that the result would be mutual military disaster with, in addition, millions of civilian deaths, even when it was assumed that the nuclear exchange had not become intercontinental. The hard fact is that there is no more reason to suppose that an exchange of nuclear fire could be restrained in a field war than there is to suppose that an intercontinental exchange of destruction could be. One could only hope.

We therefore need to ask whether the civil power could discharge its responsibilities of control in a moment of real nuclear crisis. How would the president know when to delegate the authority that Henry Kissinger is quoted as saying would have to be delegated? Daniel Ford tells us that in addition to the president, SAC and the National Military Command Center in the Pentagon also hold the “codes” for a retaliatory response. But these are the military executive agencies whereby a political decision would be implemented—they are not the elected representatives of the people. In the event of the death of the president, how does authority pass, in conformity with the Constitution, to the vice-president or to the speaker of the House? How does the high military command and everyone else learn that authority has passed—and to whom?

How, indeed, in the event of a crisis—for example a first strike by the USSR—could the president consult or inform the members of Congress? And given the inadequacies of the communications network, how could the president consult his allies, and they one another? When Prime Minister Attlee flew to see President Truman in November 1950, there was time for him to express Britain’s disquiet at the rumor that America might use a nuclear weapon against North Korea in order to prevent the total disintegration of MacArthur’s forces. Today there would be no time.

And what about the disciplined dependence of the military on the political authority? It is a fact of history that the high commands exert their own pressures in times of acute political crisis. General MacArthur did not see eye to eye with President Truman, but the President had time to dismiss him. The Pentagon brought considerable pressure to bear upon President Kennedy in the Cuban crisis—but the President’s judgment was able to prevail. What is the position today? Daniel Ford quotes a former Pentagon official as saying that while Sergeant Bilko can’t start World War III today, General Bilko certainly can. Can he?

Scott Sagan, who participated in the Avoiding Nuclear War project of the Kennedy School, and who is now a member of a Joint Chiefs of Staff directorate which deals with nuclear and chemical warfare, has thrown some light on this question in an article that has been published separately.5 It deals with three nuclear alerts, those of May 1960, at the time of the first U-2 incident, of October 1962, at the time of the Cuban crisis, and of October 1973, at the time of the Middle East war, and is probably the most detailed and revealing analysis of official and now accessible information about nuclear crises that has ever seen the light of day. He tells us that there are five gradations of nuclear alert, beginning with “fade-out” and ending with the most acute, “cocked-pistol.” The three cases of alert that he analyzes were initiated at the presidential level. The Joint Chiefs who received the order in turn issued their instructions in a way that permitted a somewhat higher level of alert than the political authority intended. Base commanders then went even higher. The point Sagan is making is that the civil authority does not appreciate sufficiently the “degree to which alert authority rests in the hands of individual military commanders…who can take what they judge are necessary steps to protect their forces.” He also writes that there is no evidence that the Russians responded in 1973 by alerting their nuclear forces, and that their response in 1962, before they backed off in Cuba, was less menacing than the US response.

The McMahon Act of 1946 placed America’s nuclear arsenal under civilian control. Today it is under that of the military. This is clearly a dilution of presidential authority, however much the release of nuclear warheads may be held in check by electronic locks and by operational procedures. But again, what degree of authority is delegated to the military so that they can deal speedily with urgent crises? Stephen Meyer believes that in the USSR the KGB, which is an executive arm of the Politburo, has a part in controlling the release of Russian nuclear warheads. It would have helped if the reader were told more about the way this is done.

It would also have helped if we were told something about the exact constitutional authority of the high military command in the US. The military do not start wars. Deciding that events had so developed that military action was called for would be a political judgment, whether the country was the victim of aggression or the government decided to undertake military action in defense of a critical national interest. In the United Kingdom the Chiefs of Staff would have the responsibility to advise whether defense or attack was likely to succeed, or whether, whatever the circumstances, defeat was a certainty. Today it would be absurd to suppose that any top military adviser would encourage his political masters to go to war when success was clearly impossible or when nuclear annihilation of their country was the likely outcome. If the political authority decided to go ahead despite such advice, its military advisers would have to comply—or resign.

President Truman is said to have had on his desk a notice saying “The buck stops here.” The buck is now, to some extent, in the hands of the military, and in consequence, if Mr. Ford has not been misled, the man who might press the doomsday button could well be a general, even if, were he rational, he would know that he had no more than a marginal chance of keeping a nuclear war under control. He would also certainly know that “war” could immediately mean the deaths of hundreds of thousands, of perhaps millions, with disintegrating skyscrapers hurtling through others as they collapsed, in cities where the streets had disappeared. It would be a vision of a hundred or even a thousand Hiroshimas. When the fires died down, a radioactive breeze would soothe the dead in the cooling rubble of Washington, New York, Moscow, Leningrad, and Paris, as winter closed in. The word “war” cannot and should not be used in association with the term “nuclear.”

What the leaders of both sides therefore need to ask is what imaginable political prize would be worth such a risk. Of course it is conceivable that were the US and the USSR to use nuclear weapons against each other, one side might throw in the sponge after a few days—given that anyone was left in authority to do that—after both had suffered millions of deaths. And then what? I suppose that such a war could start as in the scenario suggested in Hawks, Doves, and Owls—with an uprising in East Germany. But who says that in the rumbling phases leading to hostilities NATO would not disintegrate; that France and the United Kingdom would not keep out of the subsequent conflict? In a world of peace, a reunited Germany is an understandable goal, but if a war that might go nuclear were started to bring union about, there would be no Germanies to unite.

President Reagan has said that a nuclear war cannot be won, and must never be fought. So have Russian leaders. Presumably neither the President nor Mr. Gorbachev would have reached his present high office if those by whom he was elected or selected had felt that their candidates were so simple as to regard nuclear war, with its inevitable consequences, as a military option to be considered seriously in some circumstances, if either had declared publicly that in order to achieve some political purpose he might risk a lightning holocaust that could bring death to scores of millions of his fellow citizens. “Better Dead than Red,” and its converse, “Better the grave than capitalism,” might have been meaningful slogans in the days when the US and the USSR possessed only small nuclear arsenals, when the degree of nuclear devastation might have been limited.

Today they have an empty ring. There is no alternative to peace between the superpowers, which is another way of stating the message of Hawks, Doves, and Owls. The most urgent issue which the United States and the USSR now face is the avoidance of war with each other. Political differences cannot be allowed to drive them there. Neither should be drawn into a catalytic or accidental nuclear war. Both should be concerned to prevent nuclear proliferation. Both have to avoid any step that could lead to the destabilization of the present state of mutual nuclear deterrence. Above all, neither can afford to commit its future to the operation of automated systems of command and control which might either be technically defective or which, by their very perfection, could foreclose the workings of human judgment.

This is the second of two articles.

This Issue

August 15, 1985